106261 research outputs found
Sort by
Data from: Radical Deamination of Primary Amines for Initiation of Controlled Polymerization
No full textPlease cite as: Driscoll, M. E.; Nicholls, B. T.; Fors, B. P. (2025) Data From: Radical Deamination of Primary Amines for Initiation of Controlled Polymerization. [Dataset] Cornell University eCommons Repository. https://doi.org/10.7298/6syt-9j54Spectral data for the compounds and polymers reported in the associated article and Supporting Information. Data types include NMR, GPC, MALDI, and UV-Vis.National Science Foundation: CHE – 2203758
National Science Foundation Graduate Research Fellowship Program DGE – 2139899
National Science Foundation Major Research Instrumentation Program: CHE-1531632.
National Science Foundation Materials Research Science and Engineering Centers: DMR-1120296
Data from: Grab-n-Go: On-the-Go Microgesture Recognition with Objects in Hand
No full textPlease cite as:
Chi-Jung Lee, Jiaxin Li, Tianhong Yu, Ruidong Zhang, Vipin Gunda, François Guimbretière, Cheng Zhang. (2025) Data from: Grab-n-Go: On-the-Go Microgesture Recognition with Objects in Hand. [dataset] Cornell University Library eCommons Repository. https://doi.org/10.7298/7kbd-vv75.These files contain data supporting all results reported in Lee et. al. Grab-n-Go: On-the-Go Microgesture Recognition with Objects in Hand. As computing devices become increasingly integrated into daily life, there is a growing need for intuitive, always-available interaction methods — even when users’ hands are occupied. In this paper, we introduce Grab-n-Go, the first wearable device that leverages active acoustic sensing to recognize subtle hand microgestures while holding various objects. Unlike prior systems that focus solely on free-hand gestures or basic hand-object activity recognition, Grab-n-Go simultaneously captures information about hand microgestures, grasping poses, and object geometries using a single wristband, enabling the recognition of fine-grained hand movements occurring within activities involving occupied hands. A deep learning
framework processes these complex signals to differentiate among 30 distinct microgestures organized into 5 grasping poses. In a user study with 10 participants and 25 everyday objects, Grab-n-Go achieved an average recognition accuracy of 92.0%.
Furthermore, a follow-up study expanded our dataset to include an additional 10 deformable objects, and we will release this enriched dataset to the research community. These results underscore the potential of Grab-n-Go to provide seamless, unobtrusive interactions without requiring modifications to existing objects.This project was supported by the National Science Foundation Grant No. 223956
An Exceptionally Hostile Terrain: Pathways and Barriers to Strippers’ Collective Autonomy
No full text280 pagesThis thesis investigates the complex interplay of state policies and industry practices that shape the working conditions and organizing efforts of strippers in the United States. Utilizing Martha Fineman's vulnerability theory, this study examines how legal and regulatory frameworks can aid or exacerbate the precariousness of strippers' labor, while also highlighting their resilience and agency in advocating for improved working environments and societal respect. Through detailed industry research and qualitative data analysis, the research reveals the impacts of stripper resistance and organizing, offering insights into the challenges they face and their visions for a more just and equitable future. Key findings include the identification of systemic External and Internal barriers to collective autonomy and strippers’ Active Resistance and Adaptive Resilience strategies used to resist exploitation, challenge oppression, and create new pathways for community, autonomy, and voice in the workplace.2027-06-1
Combatting ill-conditioning and heavy-tailed noise in non-convex optimization
No full text170 pagesIll-conditioning is a major challenge for optimization with first-order methods. This is especially the case for stochastic optimization, where the classical preconditioners - designed to compensate ill-conditioned curvatures - are hard to construct due to the inherent variability of stochastic gradients. Compounding this challenge, heavy-tailed noise further corrupts the reliability of gradient-based update directions. This thesis confronts these challenges through three algorithms. For deterministic non-convex problems, we analyze a subgradient method whose convergence rate is independent of the functional conditioning. In stochastic regimes, we derive a stochastic gradient method with block-coordinate stepsizes that effectively combat imbalanced gradients and inhomogeneous noise. Finally, we propose a statistically efficient data-driven method that adapts to the heavy-tail distribution of the noise. In Chapter 2, we analyze a preconditioned subgradient method for optimizing composite functions , where is a locally Lipschitz function and is a smooth nonlinear mapping. We prove that when satisfies a constant rank property and is semismooth and sharp on the image of , the method converges linearly. In contrast to standard subgradient methods, its oracle complexity is invariant under reparameterizations of . In Chapter 3, we consider stochastic approximation methods with block-coordinate stepsizes and propose adaptive stepsize rules that aim to minimize the expected distance of the next iterate from an optimal point.These stepsize rules use online estimates of the mean and second moment of the search direction along each block coordinate. The popular Adam algorithm can be interpreted as using a particular heuristic for such estimation. By leveraging a simple conditional estimator, we derive variants that require fewer hyperparameters and optimizer states but obtain comparable performance. In addition, our convergence analysis relies on a simple aiming condition that is weaker than convexity, thus has broader applicability. In Chapter 4, we develop statistically efficient data-driven methods that can adapt to the heavy-tail distribution of the noise and obtain optimal performance. Our methods build upon the pioneering work of Polyak and Tsypkin in the 1980's, who developed a framework of nonlinear stochastic gradient methods where the stochastic gradients first go through a monotone nonlinear mapping, determined by the probability distribution of the noise, before being used to update the optimization variable. These methods are statistically efficient, in the sense that their asymptotic performance achieves the Cram\'er-Rao lower bound. We extend the framework of Polyak and Tsypkin to the non-asymptotic regime, showing that the same nonlinear mapping obtains the optimal statistical performance with finite samples; in addition, we allow the nonlinear mapping to be non-monotone, thus can handle heavier-tail noises than Laplacian. Empirically, we identify families of heavy-tail distributions that not only give good approximations in practice but also admit simple nonlinear mappings and can be implemented efficiently. Chapter 5 documents future directions beyond the scope of this thesis
DISCRETE GREEN’S FUNCTION BASED PREDICTIVE FRAMEWORK FOR TRIPLY PERIODIC MINIMAL SURFACES
No full text74 pagesTriply Periodic Minimal Surfaces (TPMS) are advanced porous geometries known for their high surface area, thermal efficiency, and tunable transport properties, making them ideal for thermal applications such as heat exchangers, energy storage, and catalytic systems. Accurate prediction of heat transfer in TPMS is essential for design optimization, as traditional methods like Computational Fluid Dynamics (CFD) are computationally expensive and impractical for iterative design. The Discrete Green’s Function (DGF) method, known for its computational efficiency through linear superposition, has been primarily applied to simpler geometries due to its dependence on analytical solutions or CFD-derived metrics. This work presents a novel framework that integrates DGF with Signed Distance Function (SDF)-based geometry representation, enabling mesh-free and rapid heat transfer evaluation in TPMS. By voxelizing the SDF, extracting geometric properties slice-wise, and constructing a convection-based DGF matrix via local temperature perturbations, the method allows efficient computation of heat flux and convective coefficients. This SDF-driven DGF approach retains physical accuracy while eliminating CFD dependence, offering a scalable and robust solution for rapid thermal analysis and optimization of complex porous media
ROBUST SEPAL DEVELOPMENT IS ASSOCIATED WITH UNCORRELATED GROWTH FLUCTUATIONS AND DYNAMIC MICROTUBULES
No full text169 pagesHeterogeneity exists in biology, which poses a challenge because often outcomes or responses need to be robust despite this heterogeneity. This challenge is of particular importance during morphogenesis because organ size and shape are important for their function. For example, organs need to scale to be the correct size for the organism, whether that organism is a mouse or an elephant. Organ shape can also allow for specialized function, like how beaks of birds are specialized for their diet. The size and shape of sepals, which are the four leaf-like organs on the outside flowers, allows them to enclose and protect the developing floral organs until the flower blooms. The sepals of Arabidopsis thaliana have uniform size and shape within a flower and between flowers despite heterogeneity in cell growth rates. Thus, heterogeneity is averaged spatially in the tissue and over developmental time to result in a uniform or robust outcome, and this process can be referred to as spatiotemporal averaging. The ftsh4-5 mutant has decreased heterogeneity in cell growth rates, a loss of spatiotemporal averaging of growth heterogeneity, and variable final sepal size and shape. This implies that heterogeneity in cell growth rates is necessary for robust sepal development rather than a challenge to overcome. Here, I study the organ-scale mechanism of spatiotemporal averaging through live imaging and quantitative analysis. I find that fluctuations in cell growth rates that are uncorrelated in time and space are necessary for growth to accumulate or average evenly over time and space. Decreased growth heterogeneity with correlated fluctuations, as in ftsh4-5, causes growth to accumulate unevenly, which is a loss of spatiotemporal averaging. Previously it was found that the ftsh4-5 phenotype is caused by elevated reactive oxygen species (ROS), and that microtubules are also involved in growth heterogeneity. Therefore, I examine the relationships between ROS, microtubules, and growth heterogeneity. We find that ROS inhibits microtubule dynamics and growth heterogeneity. Together, this identifies more factors involved in growth heterogeneity, and that growth heterogeneity must have uncorrelated fluctuations in order to accumulate evenly and robustly into a uniform size and shape. Altogether, my research provides insight into how heterogeneity is averaged in biology so that outcomes are robust
Research on Optimization of Delivery Routing with Time Window for Enhancing Economic and Sustainable Goals Using an Improved PSO-SA Algorithm
No full text61 pagesThis paper addresses two significant challenges facing the e-commerce logistics sector: optimizing next-day delivery routing with strict time windows and determining the optimal fleet composition of electric and diesel vehicles. We introduce a mixed-integer linear programming model for the Next-Day Green Hybrid Vehicle Routing Problem with Time Windows (NDGVRP-TW), simultaneously aimed at reducing carbon emissions, enhancing profitability, and improving customer satisfaction. Three fleet configurations—pure electric, pure diesel, and hybrid—are comparatively analyzed to identify the most profitable and environmentally sustainable composition. To effectively address this NP-hard problem, we developed a modified Particle Swarm Optimization algorithm integrated with Simulated Annealing (mPSO-SA). This algorithm uniquely incorporates random reshuffling of personal best positions and perturbation steps of the global best solution, mitigating premature convergence commonly observed in traditional optimization approaches. Computational experiments utilizing real-world data demonstrate that the mPSO-SA algorithm significantly surpasses conventional Genetic Algorithms (GA) and standard Particle Swarm Optimization (PSO), achieving over 25% improvements in convergence speed and solution quality. Our results reveal that hybrid fleets substantially outperform homogeneous fleet configurations, achieving considerable operational advantages including an 18.8% reduction in labor costs, a 26.3% decrease in distance-related expenses, and an impressive 77.7% reduction in carbon emissions compared to pure diesel fleets. Additionally, the hybrid approach notably decreases late-delivery penalties by 28.6% relative to pure electric fleets, effectively balancing the environmental benefits of electric vehicles against their operational limitations, particularly their constrained driving ranges. Our sensitivity analyses generate important managerial insights, emphasizing that fleet managers should strategically align vehicle capacities to prevent route inefficiencies, proactively adapt routing strategies in response to increasing operational costs, and meticulously schedule deliveries to minimize late penalties. Implementing these insights facilitates substantial improvements in profitability, resource optimization, and customer satisfaction compared to conventional benchmarks, providing enterprises with actionable, data-driven strategies for sustainable competitive advantage
THE BEAUTY OF THE FIELD: WEEDS IN THE MEDIEVAL IMAGINATION AND LANDSCAPE
No full text220 pagesWeeds are a difficult category to define. Plants that might be considered unwanted weeds in one context could be considered desired wildflowers in another. This project questions the present day assumptions that weeds have been universally seen as "bad" plants and that weed populations were a static, unchanging issue before the advent of modern agriculture. Studying Latin and English sources from the ancient and medieval world provides new insight into how human-weed relationships changed over time in Western Europe, while synthesizing data from archaeology, history, and palaeoclimate studies poses new questions related to long-term weed biodiversity change. Chapter 1 traces attempts to define the term "weed" from the 15th century to the present day, arguing ultimately for a definition of "unintended plants". In Chapter 2, I argue that a more relational attitude towards weeds that focused on their role in agroecosystems existed in antiquity and the early medieval period before being gradually replaced by a specifically Christian point of view in which weeds were seen as punishment for and a reminder of original sin. The second half of the dissertation turns to the contextualization and study of weeds in agroecosystems in northern Britain throughout the medieval period. Chapter 3 synthesizes palaeoclimate studies with archaeological and historical narratives related to political change and agricultural transformation to identify questions about these changing systems that I attempt to answer in Chapter 4. In this final chapter, I demonstrate through an archaeobotanical study of plant remains from Bamburgh Castle in northern England that weed biodiversity has been declining for over a millennia. I argue this biodiversity decline is likely due to landscape simplification through the subdivision of multiple estates starting in the centuries immediately preceding the Norman Conquest of 1066. Ultimately, this project shows that human relationships to weeds have always been complex, nuanced, and multi-faceted, and understanding this long history is key for understanding biodiversity loss and barriers to implementing ecological weed control measures in the present day.2027-06-1
EXPLORING ANTIBODY-VIRUS INTERACTIONS: FROM DEFINING REPERTOIRES TO RECEPTOR CHARACTERIZATION.
No full text213 pagesUnderstanding the mechanisms underlying antibody-mediated immunity and viral entry is critical for developing effective antiviral strategies and vaccines. The studies discussed in this dissertation integrates findings from recent investigations into canine parvovirus (CPV) and influenza A virus, focusing on the molecular and structural determinants of host immune responses and viral infectivity. To elucidate the antibody response to CPV, a canine-adapted B-cell cloning strategy was employed to isolate and characterize monoclonal antibodies from an immunized dog. Fluorescence-activated cell sorting (FACS) and in vitro culture enabled the identification of CPV-specific B cells, followed by amplification and cloning of immunoglobulin genes. Three monoclonal antibodies exhibited broad reactivity against CPV variants, with two sharing an identical heavy chain, and binding the B-site of the viral capsid, while the third targeted the A-site. Cryo-electron microscopy (cryoEM) analysis revealed significant epitope overlap with the viral receptor (transferrin receptor type-1, TfR) binding site. However, the antibodies displayed varying neutralization activities against CPV infection, consistent with their ability to compete for the receptor. The monoclonal antibodies here corresponded to some of the structures observed in the cryoEM analysis of polyclonal sera, including those present in a different dog than the monoclonal source. These results demonstrate that CPV infection elicits a focused antibody response, with a limited number of dominant antibodies contributing to viral neutralization and host protection. Beyond antibody-mediated immunity, viral entry mechanisms remain a key determinant of infection. Influenza A viruses (IAVs) engage surface sialic acid (Sia) moieties on host glycoproteins for cell entry, but the role of associated carrier proteins remains incompletely understood. To aid the exploration of IAV entry pathways in greater details, an artificial receptor was engineered by fusing an anti-hemagglutinin (HA) single-chain variable fragment (scFv) to an ‘ectodomain-less' TfR. When expressed in Sia-deficient cells, this receptor facilitated virus binding, endocytosis, and infection at levels comparable to Sia-expressing cells. Lower-affinity receptor variants enhanced viral uptake, and clathrin-mediated endocytosis was identified as the primary internalization pathway. These findings provide evidence that influenza A virus can utilize alternative receptor-mediated entry mechanisms independent of Sia, and could be beneficial in teasing apart IAV’s internalization mechanisms. Collectively, these studies deepen our understanding of antibody-virus interactions, viral neutralization mechanisms, and host receptor utilization. They elucidate the evolutionary dynamics between viruses and antibodies, the specificity and diversity of antibody responses to viral antigens, and the intricate interplay between viral antigens, antibody binding, and host receptor interactions
How does the land transfer policy improve China’s agricultural production efficiency?
No full text43 pagesThe purpose of this thesis is to investigate the improvement of China’s agricultural production efficiency after the reformed land transfer policy. Using panel data from 24 provinces spanning four decades, a Trans-log production function is estimated to assess the relationship between agricultural output value and four key inputs: land, labor, machinery, and fertilizer. The analysis is conducted for two distinct periods—before and after the 2014 policy reform—and incorporates fixed effects to control for time- and region-specific factors. Additionally, a heterogeneity analysis explores regional differences across eastern, central, and western China. Results show a significant shift post-reform: while the marginal contributions of individual inputs decline, production efficiency improves through stronger complementarities and better input coordination. These findings emphasize the role of land consolidation in enhancing resource allocation and provide evidence-based insights for future rural and agricultural policy design